UTFPR at SemEval-2019 Task 5: Hate Speech Identification with Recurrent Neural Networks

Proceedings of the 13th International Workshop on Semantic Evaluation (SemEval-2019), pages 519–523 Minneapolis, Minnesota, USA, June 6–7, 2019. ©2019 Association for Computational Linguistics 519 UTFPR at SemEval-2019 Task 5: Hate Speech Identification with Recurrent Neural Networks Gustavo Henrique Paetzold1, Shervin Malmasi2, Marcos Zampieri3 1Universidade Tecnol´ogica Federal do Paran´a, Toledo-PR, Brazil 2Harvard Medical School, Boston, United States 3University of Wolverhampton, Wolverhampton, United Kingdom ghpaetzold@utfpr.edu.br Abstract In this paper we revisit the problem of au- tomatically identifying hate speech in posts from social media. We approach the task us- ing a system based on minimalistic compo- sitional Recurrent Neural Networks (RNN). We tested our approach on the SemEval-2019 Task 5: Multilingual Detection of Hate Speech Against Immigrants and Women in Twitter (HatEval) shared task dataset. The dataset made available by the HatEval organizers con- tained English and Spanish posts retrieved from Twitter annotated with respect to the presence of hateful content and its target. In this paper we present the results obtained by our system in comparison to the other entries in the shared task. Our system achieved com- petitive performance ranking 7th in sub-task A out of 62 systems in the English track. 1 Introduction Abusive and offensive content such as aggression, cyberbulling, and hate speech have become per- vasive in social media. The widespread of offen- sive content in social media is a reason of concern for governments worldwide and technology com- panies, which have been heavily investing in ways to cope with such content using human modera- tion of posts, triage of content, deletion of offen- sive posts, and banning abusive users. One of the most common and effective strate- gies to tackle the problem of offensive language online is to train systems capable of recognizing such content. Several studies have been published in the last few years on identifying abusive lan- guage (Nobata et al., 2016), cyber aggression (Ku- mar et al., 2018), cyber bullying (Dadvar et al., 2013), and hate speech (Burnap and Williams, 2015; Davidson et al., 2017). As evidenced in two recent surveys (Schmidt and Wiegand, 2017; For- tuna and Nunes, 2018) and in a number of other studies (Malmasi and Zampieri, 2017; Gamb¨ack and Sikdar, 2017; ElSherief et al., 2018; Zhang et al., 2018), the identification of hate speech is the most popular of what Waseem et al. (2017) refers to as “abusive language detection sub-tasks”. This paper deals with the hate speech identi- fication in English and Spanish posts from so- cial media. We present our submissions to the SemEval-2019 Task 5: Multilingual Detection of Hate Speech Against Immigrants and Women in Twitter (HatEval) shared task. We participated in sub-task A which is a binary classification task in which systems are trained to discriminate between posts containing hate speech and posts which do not contain any form of hate speech. Our ap- proach, presented in detail in Section 4, com- bines compositional Recurrent Neural Networks (RNN) and transfer learning and achieved compet- itive performance in the shared task. 2 Related Work As evidenced …

Content truncated to 500 words